People who have understood the previous parts well and built things carefully, could skip these paragraphs because their charger should work right away. That is convenient for me since I feel unable to write an accurate, concise manual on testing and troubleshooting yet. I haven't seen enough trouble to know what parts are likely to get damaged, what errors builders, operators and users are most likely to make etc. So to people who have to refer to these paragraphs because their charger does not function properly: Please write me about what was wrong, what mistakes you have made, how you found out, how you solved it etc. I have learned so much from designing the firefly system that I myself wouldn't make many of the errors other people might make. So I need to learn from your errors if I ever want to write a useful manual on testing and troubleshooting.

The tests that are given in the subsequent paragraphs are meant to get the charger to work while in the field. So the troubleshooting guide is geared towards solving those problems that can be solved in the field. More serious problems are only identified.

The indicator on the switchboard is one of the more complicated parts of the system and if there is any fault in the system, it could very well be in the indicator part. For the moment, the indicator is less important since it is not essential for having the charger to produce electricity. So if it does not work, there is little sense in trying hard to repair it. As long as the indicator is not checked, it can not be assumed to be reliable, so it can not be used to measure whether other parts work properly. Use a tester instead.

After going through all the effort to build a charger, find a test site, bring all components there, maybe even build a weir and dig a canal, one must be quite eager to see it work for real. Very understandable, but there is a risc that once the charger, switchboard and battery are connected, one little component blows up because of a wrong connection. This could mean a lot of trouble for finding out what happened, what part was destroyed and get a spare one. That is why it is better to check everything very carefully according to the procedure of par. 4.11.2:

• If all parts weren't checked very carefully after building, it is quite likely that one will find a few minor (or major errors). Such errors could lead to other parts getting destroyed if not found in time.
• For unexperienced people, it is easier to check eveerything according to a prescribed procedure than to find the error(s) when the charger doesn't function during testing.
• It is very nice if the charger would just work once the water comes running through the turbine, the alternator reaches the right speed and a few switches are set `on'. For the one who build it, it means a boost to his or her self-confidence. And for the project as a whole a good start might make potential users will have more confidence in this new technology.

The charger is designed to charge batteries but when it is for testing and demonstrating only, also car bulbs can be used to dissipate the electricity generated by the charger. Instead of a battery, a make-shift piece of wiring with enough car bulbs to take up the output power of the charger (100 - 165 W) can be connected to the battery cable of the switchboard. Each bulb should have its own switch. There is no need to include a fuse (comparable to the battery fuse) in this circuit. To reduce the number of bulbs needed, 3 headlight bulbs and 3 bulbs of 15 W can be used (headlight bulbs of which the `low beam' filament has burnt out, are perfectly suitable). The `high beam' of headlight bulbs is ca. 55 W each so the total power of all 6 bulbs would be approximately 210 W. Using car bulbs instead of a battery has the following advantages:

• The car bulbs are cheaper than a battery.
• The battery needs to be discharged before it can be used for testing the charger (the charger is made such that it can not overcharge batteries at a high current). However, a discharged battery should not be stored for long, so if the charger turns out not to work, it has to be charged in another way.
• For demonstrating to people without technical training, it is more visible what the charger does: It produces enough electricity to make so many car bulbs burn. This is more convincing than seeing an indicator needle move and being told that the battery is being charged with so many `amperes'.
• With car bulbs, there is very little risc that vital parts are destroyed because of a wrong connection.

However, there are also additional difficulties to overcome: Without a battery, the charger probably does not start on its own (although with a mechanical regulator and a high head so that it runs fast, it might just start). For generating electricity, it needs a field current that is normally taken from the electricity that it produces itself: A chicken-and-egg situation.

To overcome this, dry cell batteries (`penlight', size `AA' will do) can be used to provide the initial field current to get it started. These starting batteries could be connected at many points in the electrical circuit. A very reliable way is:

• For use with an electronic voltage regulator, connect two or more penlight batteries in series to form a handy block. With a mechanical voltage regulator, one battery will be enough.
• Fit a thin wire directly to the `E' connection on the connector block in the switchboard.
• Fit a diode (e.g. type 1N4001) with its cathode (marked with a white band) to the `F' connection on the connector block. Fit a thin wire to the other, anode end of this diode.

Now to start, one only has to hold the wires to the block of 4 batteries for a moment. The `E' wire should go to the `-' end of the battery block but if one has the polarity reversed, nothing will happen since the diode will block any current.

When a number of lamps are switched on, the above trick might not work, at least not when only one or two batteries in series are used. So always switch off all lamps before trying to start the alternator.

When using lamps as a load, the alternator will lose its voltage if:

• Too many lamps are switched on. Then they consume nearly all electricity produced by the alternator and too little is left over to power the field.
• When a lot of lamps are switched on in one go. The moment car bulbs are switched on, they take up a much higher current because the filament inside is still cold. This also might make that there is not enough electricity left over to power the field.

So switch on lamps one by one and switch on the smaller, 15 W lamps once it seems that the charger is reaching its full capacity.

Probably still the alternator will lose its voltage many times when testing with lamps as a load. That is why an easy way of starting the alternator is very handy.

When charging a battery, there are no such problems because the battery will deliver the initial field current to get the alternator started.

In textbooks on alternators, there are strong warnings against having an alternator run without a battery being connected. In a firefly charger, there are no such problems because the alternator can never get a higher speed than say 4,000 RPM, while in a car it could be driven up to a speed of 15,000 RPM.

First check the charger itself:

• Are all bolts fixed properly. Important are the bolt that fixes the runner on the shaft, and the 3 (or two) bolts that fix the alternator to the frame.
• Does the shaft run free. If the seal has a little friction but can still be rotated easily by hand, this is no problem, the parts that touch one another will wear out in a couple of minutes once the charger runs.
• Is the penstock pipe fixed properly to the charger itself.
• Is the penstock pipe supported well. It will become much heavier once it gets filled with water.
• Is the screen at the beginning of the pipe fitted properly.
• Roughly estimate the actual head. Measure it with a spring rule or have someone stand near the charger, look from a distance and estimate how many times his/her height the top of the penstock pipe is above the charger.
  If the head is above 7 m, check whether the right blocking timber is fitted (see par. 4.1). For every 10 % extra head (above the correct value for the blocking timber), the charger will produce 15 % extra power. Fuses blowing almost immediately because the charger turns out to be too powerful is not a nice thing for your nerves during a first test. So if the head is more than expected:
  • Make a new blocking timber with the correct size for this head.
  • Place the charger a bit more up against the valley side, so that the head is as expected again.
  • Check whether there are enough spare fuses and take the risc. This is not wise if the head is say 20 % higher than expected, then the 16 A battery fuse will have to blow eventually.
  • Continue testing, but make sure that the charging current does not increase above 12 A by means of switching on less field current lamps.
• Is there any risc that water coming through the charger will erode the site. For testing, a grass cover is good enough. For longer use, there should be pebbles, stones, solid rock or boards to protect the soil underneath the charger.
• Is the switchboard located safely with regard to water splashing around.

Then some fast electrical check on the alternator:

• The brushes: Measure the resistance between `F' and `E' while turning the shaft (in the proper direction). Most of the time, it should read about 4 to 7 Ohm. If it suddenly goes up to a very high value but comes down when the shaft is rotated a little furter, there is no problem yet. But when it remains high for longer, there is something wrong with the brushes and conductor rings. The alternator might still work once it gets started. But getting it started with lamps as a load will be more difficult, so try say 4 penlight batteries in series.
  With some alternators, the brush-holder can be taken out of the back without disassembling the whole alternator. Then it is adviseable to smoothen the conductor rings and check whether the brushes move freely in their holders. Take out the brush-holder and smoothen the conductor ring by pressing a piece of fine sandpaper against it while rotating the runner by hand. Hold the alternator with the pulley end down so that any sandpaper grains will fall away from the bearing at the end of the shaft. Having them fall into the alternator itself is not a problem, they will be blown out later by the fan. Check whether the brushes move freely in the brush-holder and whether they are not worn out too much.
• Measure the voltage drop between the `E' and `A' connection, with the tester on `diode' range and the red, positive tester lead on `A'. Now the rectifier diodes are in blocking direction, see fig. 4.25, so the tester should read `overload' (it shows `OV' or its maximum value, the same value as with the tester leads not connected).
• Measure the voltage drop between `E' and `A' with the positive tester lead on `E'. Now the main diodes are in conducting direction and the tester should show a voltage drop of 0.8 to 1.0 V.
  There are two sets of diodes in series so one might expect to measure two times 0.6 to 0.8 V = 1.2 to 1.4 V. But the capacity of these diodes is so large compared to the small testing current supplied by the tester, that these diodes show a voltage drop of only 0.4 to 0.5 V each.
  If the diodes conduct in the wrong direction, maybe the tester leads have reversed polarity on `diode' range, see annex 0. Or the connections are different from what you thought.

Then one could check the electrical connections. If you are sure that all wiring has been made and checked properly, there is no need for this. But anyway it is handy to know how it could be done in case the charger doesn't work later and one has to check these.

Have the electrical circuit of fig. 4.25 at hand. For this, the charger and battery should be connected to the switchboard, but only after the battery fuse has been removed (or the wire to the battery + pole has been taken loose). This makes that the battery is connected to the switchboard only by the negative wire and it can not destroy any part.

First check the polarity of the battery: Measure battery voltage (with tester on the right voltage range, see annex 0) on the battery poles and pay special attention to the polarity:

• On the tester, the black tester lead should be in `-', `com' or something like this. The red tester lead should be in `+', `V / Ohm' or something similar.
• On the battery, the black tester lead should be on (what you think is) battery - pole and the red lead on battery + pole.
• Now the tester should display a positive voltage. Often this is just the number of volts. When there is a `-' sign before the number, clearly the voltage is negative and what you thought that was the - pole, in fact was the + pole and something is wrong...
  If you are not sure about how the tester displays a positive voltage, change the tester leads (red on - and black on +) and now the tester should display a negative voltage (with a `-' sign) when the polarity is right. Or try it out on a penlight battery, these have their polarity clearly marked.

Check the negative connections with a tester on `diode' range. If a connection exists, the tester should read 0 or near 0, and most testers produce a bleeb. Keep one test lead on the battery - pole and check with the other test lead whether the connections were made correctly:

• To `Batt.-' on the connector block
• To `D-' on the voltage regulator
• To `E' on the connector block
• To `E' on the alternator

Do not just follow a wire and then measure on the connection where it ends, read the connection codes on each component and measure on this connection. If these are O.K., check whether no wrong connections (short circuits!) were made inadvertently. So now the tester may read anything exept 0 or near 0:

• To `Batt.+' on the connector block
• To `F' on the connector block

Check the positive connections. Keep one tester lead on the lip of the fuse holder that is, via the connector and battery cable, connected to the switchboard and measure:

• To `Batt.+' on the connector block.
• To the current shunt (does not matter which side). If the tester shows no connection, probably the battery switch is in `off' position. Try out this switch and mark which position is `on', if it has not been done before.
• To `D+' on the voltage regulator
• To `A' on the connector block. If the tester shows no connection, the fuse in the fuse holder on the switchboard must be broke or missing.
• To `A' on the alternator

Check whether the following connection does not exists:

• To `F' on the connector block. On `diode' range, this will give a very low reading because `F' is connected to `E' via the field coil inside the alternator. Therefor take loose the `F' connection on the alternator during this measurement, or measure on resistance range (should give 3.5 to 6 Ohm).

Check the field connections. Keep one tester lead on the `DF' connection on the regulator and measure:

• To the `F' connection on the connector block. Now the tester will not read exactly 0 because there are the field current lamps in between. Check whether the reading drops when more lamps are switched on. Switch them off afterwards.
• To the `F' connection on the alternator (this should give the same, low value as on `F' on the connector block).

That's it for the `dry' checking. Switch `off' the battery switch and replace the battery fuse (or fit the positive wire to the battery + pole).

When testing with a battery as a load, one could check whether it supplies the field while the alternator is not running yet. Switch `on' the battery switch and check whether the field current lamps that are on, light up dimly. If they do not, rotate the alternator shaft to see whether the brushes made poor contact on the conductor rings at some points. If the lamps still don't burn dimly, either the battery was completely discharged and probably useless by now (measure battery voltage). Or there is something seriously wrong: Better switch `off' the battery switch. Look out for parts that were going up in smoke and check the fuses. Try to locate the error using the electrical circuit of fig. 4.25 and the checks of this paragraph.

If the lamps work O.K. it is worthwhile to see whether the indicator works because it is handy with further tests. Switch it to `Voltage' and check whether the voltage matches with the voltage of the battery as measured with the tester. Then switch it to `Current' and see whether it indicates a small, negative current (the battery is not being charged, but is discharged because the alternator only consumes field current). See whether this current becomes more negative if more lamps are switched on. If the indicator does not work properly, it needs to be checked and calibrated again, see par. 4.9.3, and this is a job that is best be done at ease in a workshop. For the moment, switch `off' the battery to conserve energy.

Then everything is ready for the real, wet test.

When it is sure that the charger will not be powerful enough to blow fuses and it is connected to a battery, things are easy:

1. Connect the water supply by closing the valve in the forebay tank (see fig. 4.32) or by building up a temporary wall in a canal so that the screen becomes submerged.
2. See whether enough water comes through the charger (`enough' means `a lot of', if it's the first time to see it run). If not, see next paragraph.
3. By now, the charger should start to run quite fast: About 1.7 times the optimum speed for the head given in par. 4.1. Consequently the blades of the runner produce a high-pitched sound.
4. Switch off the field current control lamps (the one that has no switch will remain on) and then switch on the battery. If the head is high, the sound will change immediately as it already begins to charge the battery.
5. Switch on more field current lamps. Once the pitch of the charger noise has dropped to about 0.6 times the original, high-pitched sound (corresponding with the free running speed), it should run close to its optimum speed and the charger should produce its maximum power.
   If the sound does not react to more field current lamps being switched on, probably the battery is nearly or fully charged and the regulator is limiting the voltage, see par. 4.9.4. You will have to discharge the battery a little before the charger can charge it at its full capacity.
   Whenever the charger suddenly starts running free, check the battery fuse first (and take care not to burn your fingers on it).

When the charger could be too powerful for the fuses, step 1 to 4 are the same as above, but step 5 should read:

5. Measure the charging current using one of the methods described further on in this paragraph. Switch on more field current lamps until the sound drops to 0.6 times its original sound, but make sure that the current does not increase above 12 A.

When the charger is connected to a set of car bulbs (instead of a battery), there should be no risc of blowing fuses since there is only the 25 A switchboard fuse in the circuit. Then step 4 and 5 should read:

4. Switch off all car bulbs and switch on all field current lamps. Then provide the initial field current using penlight batteries as described in par. 4.11.1. This should make the charger start generating electricity.
5. Measure the voltage (with the switchboard indicator or with a tester connected to `batt.-' and `batt.+' on the connector block), it should be 14 V or above since the regulator is limiting the voltage (it won't be exactly 14.7 V since the regulator can not work accurately without the dampening effect of a battery). Carefully switch on car bulbs one by one until the voltage drops to 13 V or below. The sound will drop in pitch as the alternator has to generate more power. When the pitch drops below 0.6 times that of the original sound, try switching off a field current lamp. In this way, the optimum number of car bulbs and the optimum number of field current lamps should be found at the same time.
   Switching on one 15 W car bulb too many or switching off one field current lamp too many will make the voltage collapse and the charger run free again. Nothing has been damaged, it just needs starting again following step 4.
   The car bulbs could be blown up if the charger is producing more than 14.2 V for a long time, so avoid this by switching on more car bulbs.

Hopefully everything went well and by now, the charger is running either at (nearly) maximum power or at a power that is limited to approximately 165 W (because of the 12 A limit in order to protect the fuses). If it does not, see par. 4.11.5.

While it is running for the first time, it is adviseable to keep a close watch:

1. Check whether any connections get exeptionally hot. If there is a funny, burning plastic-like smell, switch the battery switch off and locate where it comes from. Feel whether the battery connector, battery fuse holder, fuse holder on the switchboard and the connections on the alternator get hot. If one of them does, it means there is a bad connection. If this happens with the connections inside a fuse holder, the fuse will get too hot and it will blow at a current well below its rated current. Try to improve bad connections by cleaning, tightening, using `electric contact spray' or grease etc.
2. Check whether the seal around the shaft works properly: Is the alternator compartment completely dry or is there a fine mist of water coming from the fan.
3. Check the temperature of the alternator, especially of the ring of iron in the middle (this is the stator iron). As long as it remains below say 60 °, there is no problem.

The next thing to do is to make some measurements. Make notes of how the charger was running at the time (no. of field current lamps etc), how you made measurements (e.g.: With the tester or with the switchboard indicator) and the results. There is no need to do all measurements, or do them in the same order as they are described below. Just measure what you would like to know.

Then for people who aren't that experienced with testers: Testers have to be switched to the right range. A range is defined by:

1. What you want to measure: a voltage, current, resistance, a voltage drop (with the `diode' range).
2. With voltage and current ranges: Whether it is a `DC' (decent current)' value or `AC' (alternating current) value. For the charger, almost always DC measurements have to be made.
3. The maximum value for this range, say 2 V, 20 V or 200 V. Digital testers can stand tremendous overvoltages on all ranges exept the current ones. (for analogue testers, with a needle, an overvoltage of more than say 5 times the maximum value set by the range, could become dangerous). Professional digital testers have `autorange' feature: The tester itself selects the right range with the right maximum value once it is set to what you want to measure, a DC voltage, resistance or whatever.

If with a voltage or current measurement, the tester displays a value at the moment it is connected, but the reading goes to 0 again within a few tenths of a second, most likely you are using the `AC' (alternating current) range instead of the `DC' (decent current). So then put it on the right `DC' range and try again.

Voltage measurements: These are needed for the following purposes:

• To check the voltage range of the switchboard indicator so that, if it turns out to be correct, it can be used later on instead of the tester. If the indicator reading is not correct, write down both the reading of the tester and that of the indicator as a help for recalibrating it later on.
  For this measurement, the tester leads should be on the `batt.-' and `batt.+' connections on the connector block. This is practically the same as measuring over the points where the tiny wires of the voltage range of the indicator itself are connected (see the arrow's `V<sub>i</sub>' and `-V_i' in fig. 4.25).
• To measure the output voltage, which will be used later in calculating the output power. As the output voltage is taken the voltage as measured by the switchboard indicator (this makes it possible to measure it without a tester once the indicator is checked) so it should also be measured between `Batt.-' and `Batt.+'.
• To check the adjustment of the voltage regulator. This can only be done when the charger is connected to a nearly charged battery (with car bulbs, the dampening effect of the battery is missing and the regulator will not work properly, with a discharged battery, the voltage regulator is not yet limiting the voltage and the charger will run at full capacity).
  For this, measure the voltage between the `D-' and `D+' connection on the regulator.
• To check the voltage drop between the battery (or car bulbs) and the point where the regulator is connected. For this, measure the voltage at the regulator (between `D-' and `D+') and substract the voltage measured over the battery (measure this directly on the battery poles). Now preferably the charger should run at full capacity, so with a discharged battery. With car bulbs, this makes little sense because then there is no battery fuse while the voltage drop over it is significant.
  By dividing this voltage drop by the charging current (see next item), the total resistance between the regulator and the battery poles is found. As discussed in par. 4.9.5.6, the total resistance between these points should be ca. 0.065 Ohm. When it is considerably more (or less) than 0.065 Ohm, measure the voltage drops over each connection, the fuse, each wire of the battery cable separately to find which part has an unexpectedly high voltage drop.

Current measurements: These can be used for:

• Checking the current range of the switchboard indicator.
• Finding the optimum number of field current lamps to be switched on. Just switch on more lamps or switch some off until you have found the setting at which the curreng is highest.
  In fact it is not essential to measure how high the current is, as long as it is possible to say at which number of field current lamps the current must be highest. That is why also battery voltage or the voltage over the car bulbs could be used for this, since at a higher current, also the voltage will be higher. When charging a battery, the voltage will not change that much and the current is the clearest indicator for the output power.
• Calculating the power output.

The easiest way to measure the charging current is by just using the switchboard indicator. But as long as it has not been checked, it can not be relied upon. Then there are the following ways to measure the current using a tester.

• Directly with the tester. Whether this is possible depends on the tester. Some testers can measure up to 10 A continuously and up to 20 A for no longer than 10 seconds (check in its manual). Other testers might have a much lower maximum current and would be destroyed by a 12 A charging current and these can not be used in this way.
  The easiest way of connecting the tester is by holding the tester leads on the connections of the battery switch and then switching this off so that the current has to go through the tester instead of the switch. Probably the sound will change a little and the indicator on the switchboard will show a lower current because there is a significant extra voltage drop in the circuit (the current shunt inside the tester).
• Using one wire of the battery cable as a current shunt. Calculate the resistance of this one wire from its length and its cross-section:
  R = 0.0175 * l / A
  With length l and cross sectional area A in m and mm² respectively.
  Measure the voltage drop over this current shunt with the tester on mV range. Make sure that the tester leads are on the wire itself and not on a connector that might have a poor connection to the cable. Calculate the current I as I = V / R, with the voltage drop V in V (volts) and not in mV.
  Alternatively, the current shunt of the switchboard indicator could be used, if the resistance of this is known or is easy to calculate.

For checking the switchboard indicator, the direct method is more reliable. If the tester can only measure up to 5 or 10 A, the charging current can be reduced to below this maximum value by switching off field current lamps or car bulbs (roughly check this with the current shunt method). Then measure the current directly and compare it with the reading of the switchboard indicator. If the switchboard indicator is considerably wrong, write down both current readings (from the indicator and from the tester) so that you can calculate later how much longer or shorter the current shunt should be. how much longer or shorter.

The head: This is needed for comparing the power output with the expected output.

The simplest way of measuring the head is by measuring vertical distances with a spring rule or a folding rule. It does not matter whether you start at the top (at the water level around the screen) or at the bottom (at the height of the nozzle). When starting at the top, hold the top of the spring rule at the height of the water level, but so far away from the valley side that it hangs down vertically freely. Make a mark on the soil at 1 or 2 meters below this level. Climb down, hold the top at the heigth of the mark and let it hang down furter, make a new mark until you reach the nozzle. The inaccuracy of this method is in holding the top level with the mark on eyesight. It would be more accurate if the top of the rule was fixed on a straight board and the board was kept level with a carpenters level. Also the marks can be made more precise by placing sticks with the top at the correct height.

Another way is by fixing a light, tight wire from a point at the height of the nozzle to a point at the height of the water level at the top. It is best to choose such points that the thread will be quite steep. Measure accurately the angle a the thread makes with the horizontal direction. Use a protractor (a graduated arc) and a carpenters level for the horizontal. Then mark on the wire where it was fixed, remove it and measure its length l between the marks. Then the head H can be found by multiplying length l with the sine of angle a.

H = l * sin (a)

Alternatively, one could measure how much the thread climbes over 1 m of its length, and multiply this figure with the total length l of the wire.

Flow measurement: It would be interesting to know the flow also because then it is possible to calculate the overall efficiency of the charger. However, it would take quite some time to get everything set up for making flow measurements while it is not essential for getting the charger running properly. See annex 0 for how to make such a setup.

The results of the above measurements can be used for:

• Calculating the output power. This is simple: Output power P is output voltage V_i times charging current I.
• Choosing a better size for the blocking timber. Reasons for this could be:
  • The actual head might be different from what was expected, so that a new blocking timber has to be fitted anyway.
  • If power output is considerably higher than expected, the battery fuses are at risc and also the charging current is a bit too high for small capacity car batteries.
  • If power output is below expectations, it will take quite long to get large capacity solar batteries charged. Then it might be difficult to get them recharged in daytime only so that they can be used again the same night.
  • If the capacity of the canal or the creek it is connected to is lower than expected or if excess water can be used for irrigation, it might make sense to reduce the flow the charger needs. Of course then power output will be reduced as well, but that might not be such a problem.

  Before the width of a new blocking timber can be calculated, one has to choose what should become the new power output. Starting from the situation without a blocking timber, for reducing power output by say x%, x% of total nozzle width should be blocked by the blocking timber, so blocking timber width should be 51 * x/100 mm. If there was already a blocking timber of say y mm wide, one can calculate what portion of the effective nozzle width (= 51 - y mm) should be blocked as well. So then new blocking width will be: y + (51 - y) * x/100.

• Comparing the output power with the expected output power as presented in par. 4.1. If there is no blocking timber, just look up the expected output power for the measured head in fig. 4.3. When there is a blocking timber with a width Wb, usually it does nnot have exactly the right width for this head. Then look up what the expected output power would be without a blocking timber, and multiply it with (51 - Wb)/51. This ratio represents how much the flow (and with that: Output power) will be reduced by the blocking timber.

In par. 4.1, overall efficiency of the charger was estimated at 0.30. So if a charger produces say 15% more power than predicted in par. 4.1, its overall efficiency must be 15% higher than 0.30, so 0.345 (assuming that it consumes exactly the flow as predicted in par. 4.1).

In par. 4.1, friction losses in the penstock pipe were not taken into account. The part of gross, hydraulical power that is lost as pipe friction losses can be estimated by comparing the actual flow with the maximum allowable flow for that pipe diameter, length and roughness, see par. 4.10.8. So assuming the charger produces exactly the power predicted in par. 4.1 but pipe losses are 15 %, the net overall efficiency of the charger itself must be 0.30 / 0.85 = 0.353

If the measured power is higher than the expected power, the charger must be functioning quite well. Apparently the alternator has a reasonably good efficiency and the turbine part was built neatly, congratulations!

When overal efficiency ends up below 0.25, better try to find out why:

• Maybe the battery was already nearly charged and the regulator was limiting the output voltage (see par. 4.9.4 for how the charger reacts if this is the case). Then the charging current is controlled by the regulator instead of by the number of field current lamps. As a result, the charger will run above its optimum speed and produce less power because of this.
  Exactly the same happens if too few lamps were switched on in case you were testing with car bulbs as a load.
• Could friction losses in the pipe be exeptionally high, check using par. 4.10.8.
• Is there a lot of water that does not pass through the runner, but leaks away through the gap between runner and the side of the nozzle, or it flies right over the runner without entering it. Maybe the nozzle can be fitted better with respect to the runner.
• Maybe the charger is not running at nearly its optimum speed. Try again different settings for the field current lamps while measuring charging current.
  When the head is quite low, it might be that even with all field current lamps on, still the field current is too low. Then it makes sense to check whether the resistance over the field coil is O.K. (see par. 4.11.2) and smoothen the conductor rings if it is too high or unstable. If this does not help, one of the field current lamp could be replaced by a wire, so that when the corresponding switch is on, the field current lamps are circumvented and the field current will be even a bit higher than with 4 field current lamps. This has the disadvantage that then field current lamps will not show any more whether there is a field current, since none of them will burn.

Troubleshooting must be quite difficult for someone with limited technical experience and without ever having seen such a thing run before. Especially when things are not going well, one is tempted to try many different ways to get it running and the results could become more and more confusing. Therefor the following advice:

• Make notes: Of what you think the problem is, why you think that this might be the problem, and what you want to try to solve it. Make notes of what you have measured and how you made these measurements.
• Use your brains, especially when a whole lot of people are watching you hopefully, and more and more impatiently as time goes by. Do not start trying different things at random in a desperate attempt to get it running.
• Take your time. When things don't look too bright, try to make some sensible measurements that should help locating where the error must be. Then stop the testing, clear up the mess, have a good meal and try to reason out what must be the cause of this problem. Maybe read relevant paragraphs of the book again or sleep upon it.

The first step is observing what is going on. At least this should tell whether the problem is in the turbine part or with the alternator part. So check or measure:

Variables the turbine needs for functioning properly:

• The actual head as experienced by the charger: This can not be measured, but the force with which the water is pressed through the runner, gives an indication.
• The actual flow: Again, this can not be measured easily, but observing the water flowing away from the charger site gives an indication.

Things the alternator needs:

• The speed of the charger: Yet again, it can not be measured easily but the sound gives a very good indication, if one has heared the charger work normally before and can still remember the sound at that time.
• The field current: This can be estimated by the brightness of the field current lamps that are on and the number of lamps that are on. The brightness is difficult to judge because when it is relatively dark, lamps appear brighter. But usually, it is enough to know whether there is any significant field current or not. If necessary, the field current can be measured with a tester on `DC current' range at the `F' connection on the connector block (take loose this connection and connect the tester leads in between).

Output of the charger:

• Output voltage, see previous paragraph for how.
• Charging current.

And in general:

• Anything that appears unusual and might be related to the cause of the problem.

The charger can operate basically in two different ways depending on whether the regulator is limiting the output voltage or not. It is important that one is able to recognise in what way it is working (see par. 4.9.4 for more details).

When the regulator is not limiting the voltage:

• The charger will produce its full charging current.
• The output voltage will be below 14.5 V.
• The field current is set by the number of field current lamps that are switched on. Consequently charger speed and charging current will react to field current lamps being switched on or off.

When the regulator is limiting the voltage:

• The charger will produce such a high charging current that the voltage as sensed by the regulator will remain constant at 14.7 V (or a slightly different voltage when it is not readjusted properly).
• The output voltage V_i as measured by the indicator or with a tester will be between 14.5 to 14.7 V, so slightly lower than the voltage V_r experienced by the regulator because of the voltage drop over the current shunt (see fig. 4.25).
• The charging current can range from only 1 A to nearly the full charging current.
• The field current is set primarily by the regulator and switching on or off a field current lamp has no effect on speed or charging current.
  Only when so many field current lamps are switched off that the voltage as sensed by the regulator drops below 14.7 V, the regulator stops limiting the voltage and acts as above, but because of the low mumber of field current lamps, with a quite low charging current.

Then there are some things that hold the middle between observing and repairing:

• Check the fuses. Whenever in doubt, measure whether they still conduct. If a fuse was blown, do not fit a new one before checking why it blew, see table 4.4.
• Check the connections. Take them loose and fit them again. Check the connections to the battery poles and the alternator. Check whether the right connections were made (check the codes with the electrical circuit of fig. 4.25) and whether the connections were made properly (whether they are strong, tight, clean etc). Check fuses and fuse holders for dirty or oxidized contact surfaces.
• Check the switches (especially the battery switch) by switching them on and off a few times. A switch might just be in the `off' position while it should be on, a cheap switch might remain in a middle position and make no contact while it appears to be on, too light a switch might be destroyed by the large charging current.
• Check the cables, especially where they are likely to be bent often (near where they are connected). Look out for loose wires and short circuits between wires. Also check portions that run over solid rock and might have been damaged by a falling stone or tool.

In this guide, problems concerning the battery being charged are not included. It is assumed that for testing, car bulbs will be used or a new, functioning battery. See annex C for how to check batteries.

Checking alternator diodes could be difficult for someone without experience and proper instruments. Instead of checking the diodes themselves as described below, one could also check whether it is likely that diodes are destroyed:

• Has the alternator been connected to a battery with the wrong polarity once? If the battery was O.K. and there were no fuses in the circuit, the diodes must have been destroyed. Even with the proper fuses, diodes might have been destroyed before the fuse blew.
• Is the alternator performing way below expectations? With destroyed diodes, the charging current will be reduced considerably. Then even with all field current lamps switched on, the speed might remain too high. Also the efficiency of the alternator itself will be reduced, so when head is high and the speed can be brought down to the optimum speed by switching all field current lamps on, the power output will be disappointing.
  This argument also works in the other way: When the charger is functioning satisfactorily, don't bother about the diodes. Maybe one or two are destroyed, but as long as the charger works, who cares.

Another reason for not trying too hard to figure out whether the diodes are O.K. is that, once it becomes clear that one or more diodes are destroyed, one will probably have to bring the alternator to a workshop in town anyway. For replacing a diode, a spare diode and a heavy soldering iron are needed so it is unlikely that can be done in the field. So when there is clearly something wrong with the alternator, one might as well bring it to a workshop and have it repaired there.

A diode is an electronic device that allows current to flow in one direction: The conducting direction (this is the direction of the arrow that can be seen in its electronic symbol). Then for a wide range of currents, there will be a nearly constant voltage drop of some 0.6 to 0.7 V over it (however, alternator diodes show a voltage drop of only 0.4 to 0.5 V when measured with a tester on `diode' range since the testing current provided by the tester is too low compared to their large capacity).

In blocking direction, a diode allows only a negligible current to flow. Then the voltage over it will be as set by the voltage it was connected to.

Inside the alternator, there is a rectifier circuit with 6 diodes, see the alternator circuit in fig. 4.25. The top row of 3 diodes are called `positive' diodes so clearly the bottom row are the `negative' ones. Together, these 6 diodes transform the A.C. (alternating current) electricity from the 3 stator phases to D.C. (decent current) electricity that can be used to charge a battery. Mind that the diodes are drawn `upside down': They are not usually conducting. Only when the voltage of a stator phase is above the voltage on `A', its positive diode will conduct. And when its voltage is below that on `E', its negative diode will conduct.

So the two connections of a diode are not interchangeable:

• The one the arrow (of its symbol) points at, is called the kathode. This end is usually marked by a black or white band all around it.
• The other end is called the anode.

Finding out which end is which is difficult with alternator diodes because the positive diodes have their kathode connected to the housing, while the negative ones have their anode connected to the housing (but even this might be reversed in some large capacity alternators). In this way, all positive diodes can be fitted into one cooling element, and the negative ones to another one.

To check this, use a tester on `diode' range. If it displays a value of 0.4 to 0.5 V (or 0.6 to 0.7 for smaller size diodes), the diode must be connected in conducting direction. Then the diode end that is connected to the red tester lead must be the anode. Connected the other way round, the tester will display `OV' (overload), or its high, maximum value (the same value as with tester leads not connected to anything. Then the red tester lead must be connected to the cathode of the diode. But:

• Check whether the tester leads are connected to the right sockets of the tester: The black one on `COM' and the red one on `V / Ohm' or something similar.
• Maybe the polarity of the tester leads is not as usual. With digital testers, the red tester lead will almost always provide the positive testing voltage on `diode' range. Analogue testers (with a needle) usually have no `diode' range and then one could use one of the `resistance' ranges instead. But then the polarity of the tester leads could very well be reversed: The black tester lead provides the positive testing voltage.
  To find out, try with a diode once that has a clearly marked cathode.

Diodes can be destroyed by:

• A too high current in the conducting direction. In a firefly system, practically the only possible cause for this is the alternator being connected to a battery with the wrong polarity. There are fuses to protect against this, but it is not sure whether these react fast enough, the diodes might be destroyed before the fuse blows.
• A too high voltage in the blocking direction. This is not relevant for the firefly system. Diodes can stand some 100 to 300 V in blocking direction while a firefly charger connected to a head of 20 m and running at free running speed, can still generate only about 60 V. An alternator in a car can run at a far greater speed and then there is a problem. That is why in text books on alternators, there are warnings against having an alternator run without being connected to a regulator and battery.
• A too high temperature. This could happen because of soldering with a heavy soldering iron for a long time. Or the diode could overheat if is not mounted properly in the cooling element.

A destroyed diode can react in two, opposite ways:

• It might conduct in both directions. Then it shows a very low voltage drop in both directions and it acts as a short-circuit.
• It might not conduct in both directions. Then it shows `overload' in both directions.

So once a diode is isolated from the circuit it is in, it can be checked with a tester on `diode' range. Then the tester will either displays `overload' if the diode is connected in blocking direction, or display the forward voltage drop in V if the diode is connected in conducting way.

The above way of checking the diodes is simple and effective, but it involves a lot of work to isolate them. Of course only one of its two connections has to be taken loose to make sure that a measurement is not influenced by the surrounding circuit. But still, the alternator probably has to be opened up to get to the diodes. The connections to the diodes are soldered and heavy, so for taking them loose, a large soldering iron is needed. Once you know that one or more of the diodes must be faulty, this is not a problem since you will have to open it up anyway and replace at least one. But for just checking whether the diodes are allright, there is a need for less time-consuming methods.

One possible method is by measuring the voltage drops over the complete set of diodes. For this, disconnect the alternator from the switchboard. Measure the voltage drop in conducting direction (red tester lead on `E' and black tester lead on `A'):

• When all diodes are O.K., the tester should read 0.8 to 1.0 V . There are two diodes in series so the tester will display twice the voltage drop over one diode.
• When the tester displays only 0.4 to 0.5 V, at least one diode must be destroyed and have a very low voltage drop.
• When the tester displays almost 0, at least one positive and one negative diodes must be destroyed and have a very low voltage drop.
• When the tester reads `overload', either all positive or all negative diodes must have been destroyed and not conducting any more, or the polarity of the tester leads is reverse, see annex 0.

For measuring in blocking direction, change the tester leads (red on `A' and black on `E').

• Now all diodes will be in blocking direction so the tester should read `overload'.
• When the tester reads nearly 0, at least one positive and one negative diode must be destroyed and have a very low voltage drop.

With these measurements, destroyed diodes that have a very low voltage drop will be found, but not diodes that do not conduct in either direction.

Japanese brands and Fiat alternators have an `N' connection to the star point of the stator coils. By measuring between `E' and `N', and between `N' and `A' in both directions, some more information can be gathered. Now the forward voltage drop should be 0.4 to 0.5 V because it is measured over one diode only.

On such alternators with an `N' connection, it is easy to check all diodes in one go while the alternator is running. Make sure that the regulator is not limiting the voltage and that the right number of field current lamps are switched on. Keep the black tester lead on `E', measure the voltage on `N' and compare this with the voltage on `A':

• If the voltage on `N' is pretty close to half the voltage on `A', probably all diodes are O.K.
• If the voltage on `N' deviates more than say 0.25 V from half the voltage on `A', assume that at least one of the diodes is destroyed.

When the voltage on `N' is nearly half the voltage on `A', there is still a change that a negative and a positive diode are destroyed and that the effects even out one another. To find out more, one could measure the `AC' voltage on the `N' connection (so with the tester switched to an AC voltage range, have the black tester lead on `E' again). If all diodes are O.K. and the charger is producing some 12 A charging current, this should read no more than say 4 V. Every thinkable fault in the diodes will make this AC voltage increase to a much higher value.

There is no fixed value of what is normal. When the charger is hardly producing any charging current, it should be considerably less than this 4 V, while for an alternator that is running at full capacity, it could go up to maybe 8 V (this could never happen in a firefly charger since the turbine is not strong and fast enough). But still this AC voltage on the `N' connection will be somewhat higher if a relatively small alternator is producing a relatively large charging current.

It makes sense to measure this AC voltage on `N' connection once with an alternator that functions well (also note the conditions under which it was operating). Then this value can be used for a normal value with all diodes functioning.

On alternators without an `N' connection, it might be possible to remove a cover so that the diodes and phase coils become accessible. Maybe the star point can be found (three equally thick copper wires connected to one another and to nothing else). Then this can be used instead of the `N' connection and the alternator can be checked in the way described above. When only the diodes are accessible, these can be used. Measure the DC voltage of each stator phase (the thick copper wires that are connected to two diodes each). Like with the `N' connection, this voltage should be very close to half the voltage on `A'. An `AC voltage' measurement on stator coils gives no sensible information.

There are many more tricks to find out whether the diodes are functioning. But for most of them, a special measuring instrument or test device is needed, or experience in interpreting the results.

index page

next